Quantitative Analysis of seizures has recently undergone rapid and considerable advances. One of the driving forces behind this progress is the algorithm developed by FHS under the auspices of NIH. This algorithm is the first capable of accurate real-time detection, quantitative analysis and short-term prediction of clinical onset of seizures. Although the progress in this field has been substantial, further improvements in seizure analysis are both desirable and feasible. We propose a large-scale validation and refinement (using an existing data base) of a new method, the 'Intrinsic Timescale Decomposition' (ITD) developed by FHS for advanced analysis of brain or other non-stationary signals. ITD overcomes conceptual and practical limitations of existing linear and nonlinear methods for analysis of epileptic seizures, by providing in real time, highly precise time- frequency-energy-waveform localization. The end result of these research efforts will be a software package enabling more advanced online prediction, more accurate detection, quantification and imaging of epileptic seizures for use in implantable or portable devices and in conventional diagnostic equipment. These advances will form the basis for rational development of novel therapies for the automated blockage or abatement of seizures, aims which we will pursue upon successful completion of the research proposed in Phase I. PROPOSED COMMERCIAL APPLICATION Software package for automated real-time detection, prediction, and quantitative analysis of epileptic seizures for use in (a) conventional diagnostic equipment and (b) implantable or portable devices for the automated early warning and blockage of seizures.